The candidate's background in cognitive neuroscience and clinical neurology has recently culminated with a fellowship in aging and dementia. It was in this setting that he developed an interest on the effect of aging on memory, and has committed himself to a career studying the presentation, clinical course, and causes of age-related memory decline. Following the memory performance of a single cohort prospectively is a powerful approach in addressing lingering questions about memory and aging. With some exceptions, however, longitudinal studies have had difficulty in documenting age-related memory decline, either because of the learning effect associated with repeated testing or because of selective subject attrition. In a series of preliminary studies the candidate has recently identified an experimental design that demonstrates age-related memory decline using longitudinal data. The first objective of this proposal is for the candidate to receive formal training in longitudinal analysis of neuropsychological and other data. The first research goal of this proposal is for the candidate to apply this knowledge to community- based elderly individuals followed prospectively, in an attempt to better document age-related memory decline. The causes of age-related memory decline remain unknown but likely include early Alzheimer's disease (AD) as well as other physiological processes that change in an age-dependant manner. The candidate has recently developed a method using functional magnetic resonance imaging (fMRI) to study elderly individuals with memory decline. Preliminary findings suggest that this method can dissociate individuals into those with an AD-like pattern of brain dysfunction and those with a separate pattern of dysfunction. The second training objective of this proposal is for the candidate to continue his training in functional imaging. The second research goal is to validate-that the fMRI method can dissociate different etiologies by following the clinical course of individuals with memory decline prospectively.